Crecimiento y convergencia: un análisis desde la teoría de grafos

Economic phenomena have been studied considering independent actors as the unit of analysis. A confirmatory case of this observation is the study of income convergence between countries. This case assumes a fallacy of composition, where the whole is not equal to the sum of its parts. The general proposition of this study is that the global economic system is a complex (interdependent) system, where the relationships between elements must be considered in order to properly measure and understand the factors that explain growth and convergence. Therefore, the use of conventional econometric methods that assume independence between elements is not appropriate. Once this assertion is substantiated, this study adapts an Exponential Random Graph Model (ERGM) to growth theory to measure convergence between countries and deal with the intrinsic interdependence of a complex phenomenon. The results consistently show that the convergence hypothesis is accepted. This result seems to confirm that the global economic system tends to generate similar incomes if policies that perpetuate income inequality between the center and the periphery of the system were not in place. This is clearly observed in dyadic-level effects where differences in investment, complexity, and profitability of the external sector play as key factors in perpetuating disparities. However, the endogenous dimension of the accumulation and convergence process is shown, which is not timeless but determined sequentially in cyclical phases of impulses and brakes. Furthermore, when analyzed in conjunction with clustering methods, the converging process between countries arises from convergent patterns within country clubs (subgraphs of countries). Furthermore, this methodology allows for the identification of a set of interdependence effects among countries that determine income convergence, such as having converged with other countries in the past. Additionally, a decrease in the probability of systemic convergence is observed during the analysis period (1960-2019). In summary, this study contributes with a novel methodology and results to discuss recurring questions in the literature on economic growth.